Strange solar array problem

Dear Fellows,

I've run into a strange solar panel problem, I would like to pick your brains to hopefully figure out what could be going wrong.

First my setup: due to east-west facing of our home, we installed 6 solar panels on the east part of the roof and 6 on the west. 6 panels on each side (let's call this a set) are serially connected to each other. The two sets are connected in parallel. The panels are approximately 500 W each. I bought the west facing ones a few years earlier, but they are not older than 5 years old. We do have a 6.8 kW inverter and 5kWh lithuim-ion battery. The system was working when I last saw it.

When I arrived our off-grid home a couple of days ago, I noticed that the inverter is reporting 0 volts from the PV during the day -- which was unusual. I disconnected the plus and minus solar cables from the inverter and directly measured the voltage using a voltmeter and I saw 0 again. The next day I climbed the roof to further inspect the issue. I discovered that the six panels on the east side are actually producing good voltage (something close to 200 V, which was reasonable given the position of the sun). But 0 was coming from the west side. I then disconnected all the east panels one by one and individually measured their voltages. 0 on all of them (one reported something like 4V but it is much lower than what it should be).

Talking to a few solar people in our town (not necessarily knowledgeable about the matter -- but they are in this business anyway), one suggested that the east side panels could have burned out the west side ones, especially if I left the system alone without any load. His argument was that when the east side panels produce more volts in the morning than the west side ones, they can burn out the west ones. Could this really be true? Isn't the purpose of parallel connections between the sets to deal with such situations. And this system worked like this for approximately 2 years -- why choose now to burn them out?

In any case, if there is a burn out, I would assume one of the diodes or panels to burn out first potentially protecting the rest in the series kind of acting like a circuit breaker. This could be wrong thinking, I know. Unfortunately, our roof has a steep slope and it is impossible for me to look at the back of the panels to check the diodes. I would have to bring them down, but this is a multi-people job that I cannot afford to do it right now.

To provide a few more clues, this was an unusually wet, rainy, and stormy winter around our off-grid home. Perhaps water got into some places that it shouldn't have? Does the fact that parallel connecting the working and non-working ones made the combined voltage 0 suggests that there is a short circuit somewhere? Do you think just leaving the system dry out and hoping for the best could bring my panels back?

Sorry for the detailed post, but I am miserable and clue to this mystery would be very welcome!

Ahmet Oguz Akyuz wrote:  
We do have a 6.8 kW inverter and 5kWh lithium battery

.....But 0 was coming from the west side. I then disconnected all the east panels one by one and individually measured their voltages. 0 on all of them.

Did you mean "I then disconnected all the WEST panels". (You said "east")
What may I ask is the brand name of your inverter? Some have a log function that could be accessed to narrow down the time and day the problem happened.

Yes, I disconnected west panels one by one. Sorry for the typo.

The brand of the inverter is Lexron - it has a 6.2 kW rating and a MPPT voltage range of 90 - 500 Volt. I think my inverter was turned off when I arrived at the site from the on-off switch.

In theory, two sets in parallel should be safe if there is one bad panel, BUT… that is if you catch it quickly and don’t let it continue to short out day after day for months.

I have four sets in parallel so I had to use a combiner box, but I think every system should have one. It has fuses for each string, lightning protection, plus diodes to prevent shaded panels from robbing power from sunny panels.

Ahmet Oguz Akyuz wrote:Yes, I disconnected west panels one by one. Sorry for the typo.

The brand of the inverter is Lexron - it has a 6.2 kW rating and a MPPT voltage range of 90 - 500 Volt. I think my inverter was turned off when I arrived at the site from the on-off switch.

If the inverter was off, what does the monitoring of the incoming current from the array?
Also, I assume to get you by, you've disconnected the faulty string and are at least getting some solar input into the system with no other problems?

Do you know if these panels have micro-inverters? If so, perhaps the whole string went into rapid shutdown mode (from your stormy weather) and need some kind of resetting?

Thanks for your responses. After going through them and doing a lot of research online, I believe I have reached a reasonable conclusion. Let me share it here -- but the confirmation will have to wait until I bring down and hopefully inspect the junction box of the panels in about a months time (this is a remote site that I do not have regular access).

To remind you my setup, I had two strings (meaning serially connected) of panes on each side of the roof; east and west. The groups of panels were bought at different times, so the east side must have had a slightly higher voltage ratings (the panels look a bit bigger too). I made the mistake of connecting each string in parallel to each other, assuming that this is the best type of connection under such an east-west configuration. The solar guys from whom I bought the panels and help me install them did not warn me of any potential problem.

After doing research, I now realize that such an east-west configuration typically requires an extra step: 1) Either blocking diodes that prevents voltage discharge between the strings (each string gets its blocking diode in series, before the strings are paralleled). 2) Separate MPPT for each string. These two solutions are clearly presented in this video: https://www.youtube.com/watch?v=-x_6OLdAods

I will ask a related question about these two options in a separate thread.

So I now hope that lack of either solution caused a heating and a malfunction, most likely a short circuit, of the bypass diodes in the failed string. Hopefully, my junction boxes are not buried in epoxy and have easy to replace bypass diodes. Will keep you updated once I fully get to the bottom of this problem.

Hi !

I would isolate the non-working string completely and test it panel by panel, starting with polarity and open-circuit voltage on each module in good sunlight. I would also test the cable run itself, because this really does sound more like a short circuit or a ground fault than a simple mismatch issue. If one string is healthy and the other collapses to 0 V, I would not assume the good string “burned” the bad one. I would first suspect a failed connector, damaged insulation, water ingress in a junction box, or a bypass diode failure causing a short.

At this point, the fastest path is to separate the faulty string, verify each panel’s voltage and polarity one by one, and then check continuity and insulation on the cables before reconnecting anything.